Virtual reality-combined textile culture knowledge visual presentation method and system

Abstract

The invention provides a textile culture knowledge visual presentation method and system combined with virtual reality, and the method comprises the steps: firstly obtaining a target textile culture data set containing the textile technology, history and material attribute information, and constructing a three-dimensional textile model set containing a plurality of textile culture element three-dimensional models based on the target textile culture data set; dynamically matching the three-dimensional model with a corresponding textile culture data unit according to a preset rule to generate a virtual reality textile culture scene, then responding to an interaction operation instruction of a user in the scene, extracting interaction behaviors and semantic features, and finally adjusting three-dimensional model display parameters according to the interaction behaviors and the semantic features. And the updated virtual reality textile culture scene is rendered and output in real time, so that vivid and interactive display of textile culture knowledge is realized.

Description

Technical Field

[0001] The present invention relates to the field of virtual reality technology, and in particular to a method and system for visually presenting textile cultural knowledge in combination with virtual reality. Background Art

[0002] Traditionally, textile culture has been showcased through museum exhibitions, books, images, and written presentations. While museum exhibitions can showcase some physical textile artifacts, limited by space and the number of exhibits, they struggle to fully capture the diversity and complexity of textile culture. Furthermore, displays are typically static, allowing viewers to understand textile culture solely through observation and simple explanations. This lacks interactivity and immersion, hindering a deep appreciation of the charm of textile culture.

[0003] Books, pictures, and text introductions are more abstract, requiring readers to use their imagination to construct scenes and processes of textile culture. This approach is difficult for readers who lack relevant knowledge background to understand and cannot provide an intuitive experience.

[0004] With the development of computer and multimedia technologies, digital display methods such as 2D animation and video presentations have gradually emerged. While these methods have enriched the display of textile culture to a certain extent, they still have limitations. 2D animation and video presentations are pre-produced content, which viewers can only passively watch. They cannot personalize their exploration and interaction based on their interests and needs, making it difficult to meet the growing and diverse needs of viewers.

[0005] In addition, existing display methods often focus on presenting information in a single dimension, such as only introducing one aspect of textile technology, history or material properties, and are unable to organically integrate this information to form a comprehensive and three-dimensional textile cultural knowledge system. Summary of the Invention

[0006] In view of the above-mentioned problems, in combination with the first aspect of the present invention, an embodiment of the present invention provides a method for visually presenting textile cultural knowledge in combination with virtual reality, the method comprising:

[0007] Acquire a target textile culture data set, wherein the target textile culture data set includes a plurality of textile culture data units, each of which includes textile process information, textile history information, and textile material attribute information;

[0008] constructing a three-dimensional textile model set based on the target textile culture data set, wherein the three-dimensional textile model set includes three-dimensional models of multiple textile cultural elements;

[0009] According to preset virtual reality display rules, each three-dimensional model in the three-dimensional textile model set is dynamically matched with the corresponding textile culture data unit to generate a virtual reality textile culture scene;

[0010] Responding to the user's interactive operation instructions in the virtual reality textile culture scene, extracting interactive behavior features and interactive semantic features in the interactive operation instructions;

[0011] The three-dimensional model display parameters of the virtual reality textile culture scene are adjusted according to the interactive behavior characteristics and the interactive semantic characteristics, and an updated virtual reality textile culture scene is generated and rendered and output in real time.

[0012] On the other hand, an embodiment of the present invention also provides a textile cultural knowledge visualization presentation system combined with virtual reality, including a processor and a machine-readable storage medium, wherein the machine-readable storage medium is connected to the processor, the machine-readable storage medium is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the machine-readable storage medium to implement the above method.

[0013] Based on the above aspects, the present invention acquires a target textile culture data set containing information on textile processes, history, and material properties, constructs a 3D textile model set, dynamically matches the 3D models with textile culture data units, generates a virtual reality textile culture scene, responds to user interaction commands and extracts features, then adjusts the 3D model display parameters to render and update the scene in real time. This deeply integrates virtual reality technology into the presentation of textile culture knowledge, providing users with an immersive and interactive experience. On the one hand, this system integrates dispersed textile culture data and presents textile culture elements in an intuitive and three-dimensional form as 3D models. This overcomes the shortcomings of traditional 2D presentation methods, which lack intuitive and comprehensive information, making the presentation of textile culture knowledge more vivid and easy to understand. On the other hand, in terms of user interaction, by responding to user interaction commands in real time and dynamically adjusting the scene display based on interaction features, it enables deep interaction between users and the virtual scene, meeting the personalized needs of different users, greatly enhancing their sense of participation and desire to explore textile culture, and improving the effectiveness and influence of textile culture dissemination. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a schematic diagram of the execution flow of the method for visually presenting textile cultural knowledge combined with virtual reality provided by an embodiment of the present invention.

[0015] Figure 2 Schematic diagram of exemplary hardware and software components of a textile cultural knowledge visualization presentation system combined with virtual reality provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0016] The present invention will be described in detail below with reference to the accompanying drawings. Figure 1 This is a flow chart of a method for visually presenting textile cultural knowledge combined with virtual reality provided by an embodiment of the present invention. The method for visually presenting textile cultural knowledge combined with virtual reality is introduced in detail below.

[0017] Step S110: Acquire a target textile culture data set, where the target textile culture data set includes a plurality of textile culture data units, each of which contains textile process information, textile history information, and textile material attribute information.

[0018] In this embodiment, the scenario is set to visualize the traditional textile culture of a specific region. Obtaining the target textile culture data set requires various methods. For example, obtaining relevant data from the long-term collection of materials by local textile culture research groups with legal authorization.

[0019] Let the target textile culture data set be set X, which is composed of many textile culture data units Y. Each textile culture data unit Y contains textile process information Z, textile history information W, and textile material attribute information V.

[0020] The textile process information Z is further subdivided. Taking the hand-weaving process as an example, Z1 represents the fiber processing process. Z1 also includes the fiber selection step Z11, where the selection criteria parameters are set to a1, such as the fiber length range standard; the fiber combing step Z12, where the combing force is set to a2 and the combing frequency is set to a3. The yarn production process is represented by Z2. Within Z2, Z21 represents the spinning step, where the spinning speed is set to b1 and the twist angle is set to b2; and Z22 represents the yarn winding step, where the winding speed is set to b3 and the winding tension is set to b4.

[0021] Textile history information W is divided into W1 origin information, where W11 represents the approximate time period of origin and W12 represents the cultural background description related to the origin; W2 development information, where W21 is the identifier of different development stages and W22 is a description of the key points of technological change in each stage.

[0022] Textile material attribute information V, for natural fiber materials, V1 represents the fiber properties. Within V1, V11 is the fiber length distribution parameter, which consists of different length ranges and corresponding proportions, such as c1 for the short fiber length range and proportion, and c2 for the long fiber length range and proportion. V12 is the fiber strength distribution parameter, which similarly has different strength ranges and proportions, such as d1 for the low strength range and proportion, and d2 for the high strength range and proportion. V2 represents the overall physical properties of the material, V21 represents the material density, and V22 represents the material flexibility index. By carefully organizing this information from various data sources, the target textile culture data set X is ultimately formed.

[0023] Step S120: constructing a three-dimensional textile model set based on the target textile culture data set, wherein the three-dimensional textile model set includes three-dimensional models of multiple textile culture elements.

[0024] After obtaining the target textile culture data set X, we proceed to construct a 3D textile model set M. This set M will contain multiple 3D models that embody textile culture elements.

[0025] Step S121: parse each textile culture data unit in the textile culture data set to obtain a parsing result; the parsing result includes a textile process step sequence, a textile history timeline, and physical property parameters of textile materials.

[0026] For example, we analyze a textile culture data unit Ya in set X. For textile process information Z, we parse the textile process step sequence A. A consists of multiple steps: A1 is the raw material preparation step, in which the fiber raw materials are cleaned and sorted, with the cleanliness parameter set to e1 and the classification standard parameter set to e2; A2 is the spinning start step, which includes the starting spinning speed set value f1 and the fiber delivery set value f2; A3 is the twisting step during the spinning process, with the twisting force variation parameter set to f3 and the twisting time parameter set to f4, etc., forming a complete textile process step sequence A.

[0027] For the textile history information W, the textile history timeline B is parsed. B is composed of different time nodes and time periods. B1 represents the origin time node, B11 is the approximate time mark of this time node, and B12 is a brief description of the cultural background of the corresponding period. B2 represents the first important development time period, B21 is the start time mark of this time period, B22 is the end time mark, and B23 is a description of the major changes in textile technology during this time period. B3 represents the second important development time period, which also has B31 start time mark, B32 end time mark, and B33 description of technological changes, thus forming the textile history timeline B.

[0028] The physical property parameters C of the textile material are parsed from the textile material attribute information V. C includes C1 fiber arrangement-related parameters, which present the fiber arrangement direction in the material in matrix form, such as C11 representing the transverse fiber arrangement direction vector and C12 representing the longitudinal fiber arrangement direction vector; C2 material density parameters, which can be subdivided into density values of different parts, such as C21 representing the density of the material center and C22 representing the density of the edge; C3 surface texture parameters, such as C31 representing the texture roughness parameter and C32 representing the texture repetition period parameter, etc., to obtain the physical property parameters C of the textile material.

[0029] Step S122: extracting textile process features, textile history association features and textile material structure features from the analysis results.

[0030] The textile process features D are extracted from the parsed textile process step sequence A. D includes D1, the morphological features of the textile tool, such as the spinning wheel, represented by D11 for the overall contour shape parameters, D12 for the key component size parameters, and D13 for the component connection parameters; D2, the textile operation trajectory features. Taking the spinning process as an example, D21 represents the spatial trajectory coordinate sequence of the yarn movement, D22 represents the trajectory velocity change function (composed of the velocity values corresponding to different time points), and D23 represents the trajectory acceleration change function; and D3, the texture features of the textile product, including D31 for the main color distribution parameters of the finished fabric surface, D32 for the texture concavity parameter, and D33 for the texture gloss parameter.

[0031] Extract textile history-related features E from the textile history timeline B. E includes E1, a historical period label; E11, a dynasty identifier; and E12, the relative position parameter of the dynasty in the historical timeline; E2, a geographical distribution coordinate represented by a series of coordinate points, such as E21, which represents the distribution coordinates of the textile culture in a certain direction and E22, which represents the distribution coordinates in another direction; and E3, a collection of cultural element symbols, such as E31, which represents a specific textile pattern symbol and E32, which represents the symbol of a traditional textile tool. Each symbol has its own unique code and meaning.

[0032] The structural characteristics F of the textile material are extracted from the physical property parameters C of the textile material. F includes the fiber arrangement direction matrix F1, which is the fiber arrangement-related parameters mentioned above, further refined into the arrangement direction vector group of fibers at different levels F11 and the spacing parameters between fibers in each layer F12; the material density distribution map F2, which consists of density values in different regions, such as the density value of the central region F21, the density value of the edge region F22, and the density variation description of the transition region between them; and the surface texture wavelength parameter F31, which represents the main wavelength value of the texture and F32, which represents the wavelength variation range parameter.

[0033] Step S123: Input the textile process characteristics into a preset three-dimensional process model generation module to generate a corresponding textile process three-dimensional model; input the textile history association characteristics into a preset historical scene generation module to generate a corresponding historical textile scene three-dimensional model; input the textile material structure characteristics into a preset material structure analysis module to generate a textile material three-dimensional model.

[0034] For example, in one possible implementation, step S123 may be implemented by the following steps:

[0035] Step S123 1: Input the textile process features into a preset three-dimensional process model generation module to identify textile tool morphology information, textile operation trajectory information, and textile finished product texture information in the textile process features.

[0036] For the textile process feature D, when identifying the morphological information of textile tools, taking the spinning wheel as an example, we focus on the D11 overall contour shape parameters, which describe the general shape of the spinning wheel, whether it is round, square or other unique shapes; D12 key component size parameters, such as the diameter of the spinning wheel, the length of the spindle, etc.; D13 component connection method parameters, which determine whether the components are connected by mortise and tenon structure or other methods.

[0037] When identifying textile operation trajectory information, the spinning process is used as the basis to analyze the D21 spatial trajectory coordinate sequence of yarn movement. This series of coordinate points records the movement path of the yarn in three-dimensional space; the D22 trajectory speed change function uses the speed values at different time points to understand how the yarn movement speed changes over time; the D23 trajectory acceleration change function clarifies the rate of speed change.

[0038] When identifying the texture information of finished textile products, the D31 main color distribution parameter of the finished fabric surface is used to determine the main colors of different areas on the fabric surface; the D32 texture concave-convex degree parameter is used to judge the degree of undulation of the fabric surface texture; and the D33 texture glossiness parameter is used to understand the light reflection characteristics of the fabric surface.

[0039] Step S1232: generating a three-dimensional tool model based on the textile tool morphology information, wherein the three-dimensional tool model includes tool size parameters, tool material parameters, and tool motion constraint parameters.

[0040] Based on the identified textile tool morphological information, such as the D11 overall contour shape parameters of the spinning wheel, D12 key component size parameters and D13 component connection method parameters, a three-dimensional tool model G is generated.

[0041] The basic outline of the 3D tool model G is determined using the D11 overall contour shape parameters. For example, if the contour is circular, this is used as the basis for constructing the general shape of the spinning wheel. The specific sizes of the spinning wheel components are set based on the D12 key component size parameters, such as setting the spinning wheel diameter to one parameter value and the spindle length to another. The connection relationships and motion constraints between components are determined using the D13 component connection method parameters.

[0042] The three-dimensional tool model G includes the tool size parameter G1, which corresponds to the various size parameters in D12; the tool material parameter G2. Assuming that the spinning wheel is made of wood, G2 can represent the material properties such as the type and texture of the wood; the tool motion constraint parameter G3. According to the actual movement of the spinning wheel, G3 specifies the rotation range of the spinning wheel, the up and down movement range of the spindle, etc.

[0043] Step S1233: generating a three-dimensional operation trajectory model based on the textile operation trajectory information, wherein the three-dimensional operation trajectory model includes a trajectory space coordinate sequence, a trajectory speed change curve, and a trajectory force change curve.

[0044] Based on the identified textile operation trajectory information, such as the spatial trajectory coordinate sequence of yarn movement D21, the trajectory velocity change function D22, and the trajectory acceleration change function D23 during the spinning process, a three-dimensional operation trajectory model H is generated.

[0045] Based on the D21 spatial trajectory coordinate sequence of the yarn motion, the yarn motion path is precisely mapped in three-dimensional space. Using the D22 trajectory velocity variation function, corresponding velocity values are assigned to different points along the motion path, creating a dynamic effect in the yarn motion. Taking into account the force variations during the actual spinning process and combining spinning process principles, a trajectory force variation curve is derived from speed and other relevant parameters. For example, factors such as friction between the yarn and the spinning wheel components and twisting force can affect force variations. The trajectory force variation curve is determined by analyzing the relationship between these factors and parameters such as speed.

[0046] The three-dimensional operation trajectory model H includes the trajectory space coordinate sequence H1, corresponding to D21; the trajectory speed change curve H2, corresponding to D22; and the trajectory force change curve H3, which is derived from the above.

[0047] Step S1234: generating a three-dimensional texture mapping model based on the texture information of the finished textile product, wherein the three-dimensional texture mapping model includes a texture color distribution map, texture concave-convex mapping parameters, and texture light reflection parameters.

[0048] According to the identified texture information of the finished textile product, such as the main color distribution parameter of the finished fabric surface D31, the texture concave-convex degree parameter D32 and the texture glossiness parameter D33, a three-dimensional texture mapping model I is generated.

[0049] Using the D31 parameter for the primary color distribution of the finished fabric surface, we assign corresponding color values to different areas of the 3D model surface to determine the color distribution of the fabric surface. Using the D32 parameter for the texture's bumpiness, we generate a texture mapping relationship to give the fabric surface a three-dimensional feel. For example, by setting different bump heights, we can simulate a realistic texture. Using the D33 parameter for the texture's glossiness, we can set the light reflection effect, adjusting parameters such as the intensity and direction of light reflected on the fabric surface.

[0050] The three-dimensional texture mapping model I includes a texture color distribution map I1, corresponding to D31; a texture bump mapping parameter I2, corresponding to D32; and a texture illumination reflection parameter I3, corresponding to D33.

[0051] Step S1235: According to a preset process dynamic combination rule, the three-dimensional tool model, the three-dimensional operation trajectory model and the three-dimensional texture mapping model are dynamically spliced to generate the textile process three-dimensional model.

[0052] Preset process dynamic combination rules dictate the assembly of the 3D tool model G, the 3D operation trajectory model H, and the 3D texture mapping model I. First, the 3D tool model G is placed in a suitable position as a foundation. Based on the actual spinning operation logic, the 3D operation trajectory model H is then associated with the 3D tool model G. For example, the yarn's trajectory (the 3D operation trajectory model H) is aligned around the spinning wheel components (the 3D tool model G), ensuring that the trajectory's starting and ending points, as well as the course of the movement, match the spinning wheel's structure and operation.

[0053] Then, the 3D texture mapping model I is attached to the final fabric model surface. Based on the final position and shape of the fabric during the spinning process, the texture color distribution map I1, texture bump mapping parameters I2, and texture illumination reflection parameters I3 are applied to the fabric model surface, giving it a realistic texture effect. Through this dynamic splicing, a complete 3D textile process model J is generated.

[0054] Step S1236: inputting the textile history association features into a preset historical scene generation module, and extracting the historical period labels, geographical distribution coordinates and cultural element symbol sets from the textile history association features.

[0055] For the textile history-related feature E, the historical period label E1 is extracted, where E11 clearly identifies the dynasty, such as the Tang Dynasty and the Song Dynasty, and E12 represents the relative position parameter of the dynasty in the historical timeline, which is used to locate the position of the dynasty in the entire historical development process.

[0056] Extract E2 geographical distribution coordinates. These coordinates consist of a series of points. For example, E21 represents the distribution coordinates of the textile culture in a certain area in the north, and E22 represents the distribution coordinates in a certain area in the south. These coordinates are used to determine the distribution of textile culture in different geographical locations.

[0057] Extract the E3 cultural element symbol set. For example, E31 represents a specific textile pattern symbol, which may be a cloud pattern symbol with regional characteristics; E32 represents the symbols of traditional textile tools, such as spinning wheels and looms. Each symbol has its own unique code and meaning, representing different cultural elements.

[0058] Step S1237: matching a corresponding three-dimensional building basic grid model from a preset historical building style database according to the historical period label, and performing geographic location spatial alignment on the three-dimensional building basic grid model based on the geographical distribution coordinates.

[0059] Based on the dynasty identifier E11 in the extracted historical period label E1, a matching 3D architectural basic mesh model M is searched from the preset historical architectural style database L. For example, if E11 refers to the Ming Dynasty, a model M with Ming Dynasty architectural style characteristics is found in database L. This model includes typical Ming Dynasty architectural features such as roof forms and wall structures.

[0060] Based on the geographic coordinates E2, the 3D architectural foundation mesh model M is spatially aligned. Assuming E21 is a specific coordinate point, the model M is adjusted to the 3D spatial location corresponding to that coordinate point, ensuring that the model's position in virtual space aligns with the actual geographic distribution of the textile culture. Through operations such as translation and rotation, the model M's orientation in space is aligned with the direction indicated by the geographic coordinates, completing the geographic spatial alignment.

[0061] Step S1238: Identify the textile tool symbols, traditional clothing symbols and decorative pattern symbols in the cultural element symbol set, and call the corresponding three-dimensional tool sub-model, three-dimensional clothing sub-model and three-dimensional pattern sub-model from the preset cultural element model library.

[0062] For the cultural element symbol set E3, the textile tool symbol is identified, such as the symbol representing the spinning wheel in E31, and the corresponding 3D tool sub-model O is called from the preset cultural element model library N. The 3D tool sub-model O has the detailed structure and appearance characteristics of the spinning wheel and corresponds to the actual spinning wheel represented by the symbol.

[0063] Identify traditional clothing symbols, such as the symbol representing traditional textile clothing of a certain period in E32, and call the corresponding 3D clothing sub-model P from the cultural element model library N. This 3D clothing sub-model P presents the style, color, texture and other characteristics of the clothing of that period.

[0064] To identify decorative pattern symbols, such as those with regional characteristics in E33, the corresponding 3D pattern sub-model Q is called from the cultural element model library N. The 3D pattern sub-model Q accurately restores the shape, color, and details of the pattern, consistent with the decorative pattern represented by the symbol.

[0065] Step S1239: According to the textile scene layout rules corresponding to the historical period labels, the three-dimensional building basic grid model, the three-dimensional tool sub-model, the three-dimensional clothing sub-model and the three-dimensional pattern sub-model are combined according to the spatial topological relationship to generate an initial historical scene model.

[0066] Based on the textile scene layout rules corresponding to the dynasty identifier E11 in the historical period label E1, the 3D building base mesh model M, 3D tool sub-model O, 3D clothing sub-model P, and 3D pattern sub-model Q are combined. For example, if E11 is the Qing Dynasty, the 3D building base mesh model M is placed in an appropriate position as the background subject, according to the layout characteristics of the Qing Dynasty textile scene.

[0067] Place the 3D tool sub-model O, such as a spinning wheel model, in a location relevant to the building and consistent with actual operating locations, such as within a room. Place the 3D clothing sub-model P, representing the clothing of Qing Dynasty textile workers, next to the spinning wheel to simulate the worker's operation. Place the 3D pattern sub-model Q, such as a model of a decorative pattern characteristic of the Qing Dynasty, on the walls of the building, textile tools, or clothing, according to the actual location of the pattern in the scene.

[0068] According to the spatial topological relationship between these models, their relative positions, front-to-back order, etc. are determined, and they are combined together through operations such as translation, rotation, and scaling to generate the initial historical scene model R.

[0069] Step S12310: Mapping the time axis parameters corresponding to the geographic distribution coordinates to the initial historical scene model, dynamically loading the ambient lighting parameters and climate simulation particle effects that match the time axis parameters, and generating the three-dimensional model of the historical textile scene.

[0070] The time axis parameters corresponding to the geographic distribution coordinates E2, such as information on different seasons and time periods, are mapped to the initial historical scene model R. For example, if the area corresponding to E2 has specific climate and lighting characteristics in a certain season, the time axis parameters are determined based on these characteristics.

[0071] Based on the timeline parameters, matching ambient lighting parameters are dynamically loaded. For example, in the morning, relatively soft, warm-toned lighting parameters are loaded, adjusting the brightness and color of the model surface to simulate the effect of morning sunlight. At noon, bright, white-toned lighting parameters are loaded to simulate the strong sunlight of noon. Simultaneously, climate simulation particle effects are dynamically loaded based on the timeline parameters. If the corresponding season is winter, a particle effect representing snowflakes is loaded, adjusting parameters such as particle size, density, and movement speed to match the actual winter snowfall scene. If it is summer rainstorm season, a particle effect representing raindrops is loaded, and the particle-related parameters are similarly carefully adjusted to simulate a realistic rainfall scene. By combining the timeline parameters with the ambient lighting parameters and climate simulation particle effects, the initial historical scene model R is further rendered and refined, ultimately generating a 3D model S of the historical textile scene.

[0072] Step S12311: inputting the textile material structural characteristics into a preset material structure analysis module to analyze the fiber arrangement direction matrix, material density distribution map and surface texture wavelength parameters in the textile material structural characteristics.

[0073] The textile material structural features F are input into the pre-set material structure analysis module T. First, the fiber orientation matrix F1 1 in F is analyzed. This matrix describes the orientation of the fibers in the material. Assume that the elements in the matrix are vectors, each representing the orientation of the fibers at a different level. For example, let vector a1 represent the orientation of the fibers in the first layer, vector a2 the orientation of the fibers in the second layer, and so on. By analyzing these vectors, the fiber orientation angles are determined.

[0074] Analyze the material density distribution graph F2, which shows the density distribution of different regions of the material. Assume that F2 consists of the density values of different regions, such as the density value of region b1 is d1, the density value of region b2 is d2, and so on. Based on these density values, adjust the vertex density of the multi-level fiber geometry.

[0075] Analyze the surface texture wavelength parameter F3 1, which determines the repetition period of the material surface texture. Let the surface texture wavelength parameter be c1, and calculate the texture repetition period based on this parameter.

[0076] For the fiber orientation matrix F11 within the textile material structure F, each element of the matrix is carefully analyzed. The matrix can be represented as a two-dimensional array, such as F11[i][j], where i represents the fiber layer and j represents the dimension of the vector describing the fiber orientation at that layer. By interpreting these elements, the orientation angle of each fiber layer's geometric structure can be determined. For example, by calculating the angle between vectors, let vector a and vector b represent the orientation vectors of two adjacent fiber layers, respectively. Using conventional vector angle calculation formulas from related technologies, the angle between these two layers can be determined, thereby clarifying the changes in fiber orientation.

[0077] The material density distribution map F2 can be considered as a collection of density values for different regions. Each region can be represented by an identifier, such as r1, r2, r3, and so on, corresponding to density values ρ1, ρ2, ρ3, and so on. By analyzing these regions and density values, we can understand the density differences in different parts of the material.

[0078] The surface texture wavelength parameter F3 1 is an important parameter for determining the surface texture characteristics. Let its value be λ, which directly affects the repetition period of the texture.

[0079] Step S12312: Generate a multi-level fiber geometric structure according to the fiber arrangement direction matrix, wherein the direction angle of each layer of the fiber geometric structure is determined by the vector angle of the corresponding layer in the arrangement direction matrix.

[0080] Generate a multi-level fiber geometry U based on the fiber arrangement direction matrix F11. Starting from the first layer of the matrix, let the vector of the first layer be v1, and determine the initial direction of the fiber geometry of the first layer based on this vector. For the second layer, let the vector be v2, and by calculating the angle θ12 between v1 and v2 (using the vector angle calculation method), determine the angle of change in the direction of the second layer of fibers relative to the first layer. Similarly, for the nth layer, let the vector be vn, and by calculating the angle θ(n-1)n with the vector vn-1 of the previous layer, determine the direction angle of the fiber geometry of the nth layer. In this way, the multi-level fiber geometry U is constructed layer by layer, and the direction angle of each layer is accurately determined by the vector angle of the corresponding layer in the arrangement direction matrix.

[0081] Step S12313: adjusting the vertex density of the multi-level fiber geometric structure based on the material density distribution map, so that the vertex spacing in the high-density area is reduced according to a preset compression ratio, and the vertex spacing in the low-density area is increased according to a preset expansion ratio.

[0082] Adjust the vertex density of the multi-level fiber geometry U based on the material density distribution map F2. For a high-density region, assume the region identifier is r_high and its density value is ρ_high. Assume the preset compression ratio is s_high. For the vertices of the multi-level fiber geometry U within this region, calculate the new vertex spacing. For example, if the original vertex spacing is d0, the new vertex spacing d_high = d0 * s_high, thereby reducing the vertex spacing and increasing the vertex density.

[0083] For a low-density region, assume the region identifier is r_low and its density value is ρ_low. Assume the preset expansion ratio is s_low. For the vertices of the multi-level fiber geometry U within this region, calculate the new vertex spacing. For example, if the original vertex spacing is d0, the new vertex spacing d_low = d0 / s_low, thereby increasing the vertex spacing and reducing the vertex density. In this way, the vertex density of the multi-level fiber geometry U is comprehensively adjusted based on the material density distribution map F2.

[0084] Step S12314: Calculate the texture repetition period according to the surface texture wavelength parameter, and scale the UV coordinates of the preset basic textile texture map based on the repetition period to generate a dynamic texture mapping relationship adapted to the multi-level fiber geometric structure.

[0085] The texture repetition period P is calculated from the surface texture wavelength parameter F31. Assuming that there is a function relationship f related to the surface texture wavelength parameter F31, such that P=f(F31), the texture repetition period P is determined by this function.

[0086] The UV coordinates of the preset basic textile texture map V are scaled based on the texture repetition period P. Assume that the UV coordinates are (u, v) and the scaling factor is k, where k is related to the texture repetition period P. The k value is determined by a conventional mapping relationship in the relevant art. For example, the new UV coordinates (u_new, v_new) = (u*k, v*k). Through this UV coordinate scaling, the basic textile texture map V is adapted to the multi-level fiber geometry U, generating a dynamic texture mapping relationship W. This dynamic texture mapping relationship W can accurately display the texture map at the corresponding position according to the morphology and surface characteristics of the multi-level fiber geometry U, presenting a realistic material surface texture effect.

[0087] Step S12315: Binding the adjusted multi-level fiber geometric structure with the dynamic texture mapping relationship, and superimposing the transmittance gradient parameter generated according to the material density distribution map to generate the three-dimensional model of the textile material.

[0088] The multi-level fiber geometry structure U with adjusted vertex density is bound to the dynamic texture mapping relationship W. This means that the texture mapping relationship W is accurately applied to the surface of the multi-level fiber geometry structure U so that the texture can be displayed on the fiber structure in the expected manner.

[0089] Generate a transmittance gradient parameter X based on the material density distribution map F2. For each region in the material density distribution map F2, let the region identifier be r and the density value be ρ. Then, using a density-dependent function g, generate a transmittance value t = g(ρ). The transmittance values of each region constitute the transmittance gradient parameter X.

[0090] The transparency gradient parameter X is superimposed on the textured multi-level fiber geometry U. For each section of the multi-level fiber geometry U, the transparency is adjusted based on the transmittance of the area in which it resides. For example, in high-density areas, due to their lower transmittance, the corresponding transparency decreases; in low-density areas, due to their higher transmittance, the corresponding transparency increases. This operation generates the final 3D textile material model Y, which not only displays the fiber arrangement structure and surface texture, but also reflects the varying transmittance effects caused by varying material density.

[0091] Step S130: According to preset virtual reality display rules, each three-dimensional model in the three-dimensional textile model set is dynamically matched with the corresponding textile culture data unit to generate a virtual reality textile culture scene.

[0092] After having a three-dimensional textile model set M, according to the preset virtual reality display rules, each three-dimensional model is dynamically matched with the corresponding textile culture data unit Y to generate a virtual reality textile culture scene Z.

[0093] Step S131: extracting the cultural display priority parameters, user preference matching parameters and device rendering performance parameters from the textile culture data unit.

[0094] Extract the cultural display priority parameter A1 from the textile cultural data unit Y. This parameter can be composed of multiple sub-parameters, such as A11, which indicates the display priority of textile craftsmanship-related content, A12, which indicates the display priority of textile history-related content, and A13, which indicates the display priority of textile materials-related content. These priority parameters can be derived by assessing the importance of each piece of information in the textile cultural data unit Y, for example, based on factors such as the uniqueness of the information and its criticality to overall cultural understanding.

[0095] Extract the user preference matching parameter A2. The user preference matching parameter A2 includes multiple aspects, such as A21 representing the user's historical interaction record features, which can be a collection of the user's past interactions with textile culture, such as the textile process steps the user has viewed and the historical periods the user has paid attention to; A22 represents the user's attention distribution features, which can be a distribution vector describing the user's attention to different textile cultural elements, such as the degree of attention paid to textile tools, fabric textures, and other elements; A23 represents the user's device operation habit features, such as A231 representing the average speed of the user's operation of the virtual reality device, and A232 representing the frequency of the user's operation.

[0096] Extract device rendering performance parameters A3. Device rendering performance parameters A3 relate to the device's ability to process graphics rendering. For example, A31 represents the device's graphics processing unit (GPU) performance indicators, which can be a set of parameters reflecting comprehensive performance such as GPU computing power and video memory size. A32 represents the device's memory performance indicators, such as memory bandwidth and memory space available for graphics rendering.

[0097] Step S132: determining an initial display position of each three-dimensional model in the three-dimensional textile model set according to the cultural display priority parameter.

[0098] The initial display position of each three-dimensional model in the three-dimensional textile model set M is determined based on the cultural display priority parameter A1. For the display priority of A11 textile technology-related content in the cultural display priority parameter A1, if the value of A11 is higher, then the three-dimensional models related to textile technology, such as the textile technology three-dimensional model J, will be placed in a more conspicuous position in the virtual reality scene. Assuming that the virtual reality scene can be regarded as a three-dimensional space, with the center of the scene as the reference point, the position coordinates of the model in the space are determined according to the priority. For example, the model with the highest priority may be placed near the center of the scene and at the height of the human eye, and the coordinates of the position are set to (x1, y1, z1); the model with lower priority may be placed at the edge of the scene or at a higher or lower position, such as the coordinates (x2, y2, z2). In this way, the initial display position of each three-dimensional model is determined based on the cultural display priority parameter A1.

[0099] Step S133: adjusting the display size, display angle, and display dynamic effect of the three-dimensional model according to the user preference matching parameters.

[0100] Step S1331: Identify historical interaction record features, user attention distribution features, and user device operation habit features in the user preference matching parameters.

[0101] The user preference matching parameter A2 is analyzed to identify historical interaction record features A21. Historical interaction record features A21 are a collection of rich user interaction information. For example, it may record information such as the number of times a user viewed a specific textile process step, such as the spinning step, and the duration of their visit. This information can be represented by a data structure, assuming a two-dimensional array. The first dimension represents different textile cultural elements, such as textile process steps and historical events, and the second dimension represents the corresponding interaction data, such as the number of views and duration of their visit.

[0102] Identify the user attention distribution feature A22. The user attention distribution feature A22 can be represented as a vector, where each dimension of the vector corresponds to a different textile cultural element, and the value of the vector represents the user's attention to that element. For example, let vector A22 = [c1, c2, c3, ...], where c1 represents the user's attention to textile tools, c2 represents the user's attention to fabric texture, c3 represents the user's attention to textile history, and so on.

[0103] Identify user device operation habit feature A23. User device operation habit feature A23 includes multiple parameters, such as A231, the average speed of the user operating the virtual reality device, which can be calculated by recording the displacement and time of the user operating the device over a period of time; A232, the frequency of user operations, which is obtained by counting the number of operations within a set time.

[0104] Step S1332: adjusting the display size of the three-dimensional model according to the historical interaction record characteristics, so that the display size of the three-dimensional model with a higher historical interaction frequency of the user is larger.

[0105] Adjust the display size of the three-dimensional model based on the historical interaction record feature A21. Taking the textile process three-dimensional model J as an example, assume that the historical interaction record feature A21 records the user's interaction frequency for different process steps such as spinning and weaving. Let the interaction frequency of the spinning step be f1, and the interaction frequency of the weaving step be f2. Using a function h related to the interaction frequency, calculate the three-dimensional model display size adjustment factor corresponding to each process step. For example, the display size adjustment factor s1 = h(f1), s2 = h(f2). For the three-dimensional model part corresponding to the spinning process, multiply its original size by the adjustment factor s1 to obtain the new display size. If the original size is (11, w1, h1), the new display size is (11*s1, w1*s1, h1*s1). In this way, the display size of the three-dimensional model with a higher historical user interaction frequency is larger, highlighting the content that the user is interested in.

[0106] Step S1333: adjusting the display angle of the three-dimensional model according to the user attention distribution characteristics, so that the viewing angle of the three-dimensional model with higher user attention is closer to the user's viewpoint center area.

[0107] Adjust the display angle of the three-dimensional model according to the user attention distribution feature A22. In the virtual reality scene, the central area of the user's viewpoint is set as a specific spatial range, which is determined by the user's position as a reference point. For a three-dimensional model, such as the three-dimensional model S of the historical textile scene, it is assumed that in the user's attention distribution feature A22, the degree of attention to the historical building elements is c_building, and the degree of attention to the character elements is c_character. According to the level of attention, calculate the rotation angle of the three-dimensional model part corresponding to each element. For example, for the model part corresponding to the historical building element, the rotation angle θ_building = i(c_building) is calculated by a function i related to the degree of attention. Rotate this part of the model around a specific axis (such as a vertical axis or a horizontal axis) by an angle of θ_building so that the perspective of the elements with a high degree of attention is more toward the central area of the user's viewpoint to meet the user's attention needs for different elements.

[0108] Step S1334: adjusting the dynamic display effect of the three-dimensional model according to the operating habit characteristics of the user device, so that the rotation speed and zoom sensitivity of the three-dimensional model match the operation delay parameters in the operating habit characteristics of the user device.

[0109] Adjust the dynamic display effects of the 3D model based on the user's device operation habit characteristic A23. Taking the operation delay parameter as an example, let the average operation delay of the user operating the virtual reality device be d. For the rotation speed of the 3D model, let the original rotation speed be v_rotate. Using a function j related to the operation delay, calculate the adjusted rotation speed v_rotate_new = j(d, v_rotate). For example, if the operation delay d is large, appropriately reduce the rotation speed to make the operation smoother.

[0110] For zoom sensitivity, let the original zoom sensitivity be s_scale. Similarly, function j is used to calculate the adjusted zoom sensitivity s_scale_new = j(d, s_scale). This way, the rotation speed and zoom sensitivity of the 3D model match the operation delay parameter in the user's device operation habits, providing a more comfortable interaction experience for the user.

[0111] Step S134: Optimizing the texture resolution, number of model faces, and lighting calculation complexity of the three-dimensional model according to the device rendering performance parameters.

[0112] The 3D model is optimized based on the device rendering performance parameter A3. For the GPU performance indicator A31 in the device rendering performance parameter A3, if A31 shows that the GPU computing power is relatively low, in order to ensure smooth rendering of the scene, the texture resolution of the 3D model needs to be reduced. Assume that the original texture resolution of the 3D model is (r1, r2), and a function k related to GPU performance is used to calculate the adjusted texture resolution (r1_new, r2_new) = k(A31, (r1, r2)). For example, if the GPU computing power is insufficient, the texture resolution can be appropriately reduced to reduce the amount of texture data and reduce the burden on the GPU.

[0113] For the number of model faces, let the original number of model faces be n. Similarly, based on the GPU performance indicator A31 and the memory performance indicator A32, the adjusted number of model faces, n_new = k(A31, A32, n), is calculated using the function k. For example, if memory bandwidth is limited, the number of model faces can be appropriately reduced to simplify the model structure and avoid memory overflow.

[0114] For lighting calculation complexity, let the original lighting calculation complexity be c. Based on the GPU performance indicator A31, the adjusted lighting calculation complexity c_new = k(A31, c) is calculated using function k. For example, if the GPU computing power is low, a simpler lighting model can be used to reduce the lighting calculation complexity and ensure that the scene can be rendered in real time.

[0115] Step S135: Arrange the adjusted three-dimensional models according to a preset spatial layout rule, and add virtual reality environment lighting effects and background sound effects to generate the virtual reality textile culture scene.

[0116] Arrange the 3D models that have undergone the aforementioned adjustments according to a preset spatial layout rule. This preset spatial layout rule can be a set of rules describing the relative positions of the 3D models within the virtual reality scene. For example, a 3D model of textile crafts, J, can be placed at the front of the scene as the primary display object; a 3D model of a historical textile scene, S, can be placed at the back as a background to provide cultural context; and a 3D model of textile materials, Y, can be placed to the side to facilitate viewing of material details.

[0117] Add VR ambient lighting effects. Choose the appropriate lighting type and parameters based on the theme and atmosphere of the scene. For example, if the scene is set as a textile workshop during the day, choose bright ambient light and set parameters such as light intensity and color. Set the light intensity to I and the color parameters to (r, g, b). Adjust these parameters to simulate real daytime lighting effects.

[0118] Add background sound effects. Choose appropriate background sound effects based on the scene. For example, in a textile workshop scene, add sounds like a spinning wheel and a loom. Set the volume of the spinning wheel sound to v1 and the volume of the loom sound to v2. By adjusting parameters like volume, the background sound effects blend in with the scene, enhancing the user's sense of immersion and ultimately generating a virtual reality textile culture scene Z.

[0119] Step S140: Responding to the user's interactive operation instructions in the virtual reality textile culture scene, extracting interactive behavior features and interactive semantic features in the interactive operation instructions.

[0120] When the user operates in the virtual reality textile culture scene Z, the system responds to the user's interactive operation instructions and begins to extract the interactive behavior features and interactive semantic features.

[0121] Step S141: Capturing the user's gesture data, eye tracking data, and voice input data in the virtual reality device.

[0122] The user's gesture data B1 is captured using the sensors and input devices of the virtual reality device. The gesture data B1 can be a data set containing information in multiple dimensions. For example, taking a common hand motion capture system as an example, it may contain a sequence of position coordinates of each joint of the hand. Assuming that the hand has n joints, the three-dimensional spatial coordinates of each joint at each time t can be expressed as (xi_t, yi_t, zi_t), where i ranges from 1 to n. These coordinate sequences record the movement trajectory of the hand in space. At the same time, the gesture data B1 may also contain hand posture information, such as the direction of the palm, the degree of extension of the fingers, etc. This information can be represented by specific parameters. For example, the direction of the palm can be described by a direction vector (dx, dy, dz), and the degree of extension of the fingers can be reflected by the bending angle θi of each finger joint, where i ranges from 1 to 5 to represent different fingers.

[0123] Capture eye tracking data B2. Eye tracking data B2 mainly focuses on the movement of the user's eyeballs and the information of the gaze point. It includes a gaze point coordinate sequence. Assuming that the screen or virtual scene space is used as the reference system, the two-dimensional or three-dimensional coordinates of the user's eye gaze point at each moment t are (xt, yt) or (xt, yt, zt). These coordinates constitute a gaze point coordinate sequence, reflecting the change in the user's visual focus position in the scene. In addition, the eye tracking data B2 also includes the gaze duration, that is, the length of time the user stays at each gaze point. Let the gaze duration for the gaze point (xt, yt) be dt. Some physiological characteristic data of the eyeball, such as the pupil diameter change curve, are also included. Let the value of the pupil diameter at moment t be pt, and the pupil diameter change curve is formed as it changes over time.

[0124] Capture voice input data B3. The voice input data B3 is first a voice signal spoken by the user, which is converted into a digital signal after being collected by the device. The digital signal is then preliminarily processed, for example, by converting it into voice text content through voice recognition technology. Suppose the voice text content is a string S, which contains the semantic information expressed by the user. At the same time, the voice emotion polarity can be extracted through sentiment analysis technology, such as judging the positive, negative or neutral emotions contained in the voice, represented by a numerical value e, which can take values within a specific range, such as from -1 (representing extremely negative) to 1 (representing extremely positive). The voice instruction type is also an important part. The voice instruction type is determined by analyzing the voice text content and matching it with the preset instruction rules. Suppose the voice instruction type is represented by an identifier c. For example, c can represent different types of instructions such as "query material properties" and "play historical explanations".

[0125] Step S142: extracting the gesture type code, gesture motion trajectory and gesture force parameter from the gesture action data to generate a gesture behavior sub-feature in the interaction behavior feature.

[0126] Extract the gesture type code C1 from the gesture action data B1. This requires analyzing the hand posture and movement pattern in the gesture action data. For example, by defining a series of gesture templates, the captured gesture actions are matched with these templates. Assume that there are m gesture templates, each template has a specific hand joint position relationship and movement characteristics. For the currently captured gesture action, calculate its similarity with each gesture template, and set the similarity calculation function to sim(B1, Tj), where Tj represents the jth gesture template, and j ranges from 1 to m. By comparing these similarity values, find the template with the highest similarity, and its corresponding identifier is the gesture type code Cl. For example, if the similarity with the "fist" gesture template is the highest, then the code corresponding to the "fist" gesture is C1.

[0127] Extract the gesture motion trajectory C2. The gesture motion trajectory C2 is mainly based on the hand joint position coordinate sequence in the gesture action data B1. Taking a key joint (such as the index finger tip joint) as an example, its coordinates (x_t, y_t, z_t) at each moment constitute the motion trajectory in three-dimensional space. In order to describe the trajectory more clearly, it may be necessary to process the coordinates, such as removing some abnormal points caused by noise. A threshold can be set to correct or eliminate points whose distance from adjacent points exceeds the threshold. Assuming the distance threshold is d, for a coordinate point (x_t, y_t, z_t), if the distance sqrt((x_t-x_t-1)^2+(y_t-y_t-1)^2+(z_t-z_t-1)^2) between it and the adjacent point (x_t-1, y_t-1, z_t-1) is greater than d, then the point is processed. The processed coordinate sequence is the gesture motion trajectory C2.

[0128] Extract the gesture force parameter C3. The gesture force parameter C3 can be obtained in a variety of ways. For example, some virtual reality devices are equipped with pressure sensors that can directly measure the force applied by the hand when making gestures. Let the pressure value of each part of the hand measured by the sensor be pi, and i from 1 to k represents different parts of the hand. These pressure values can be converted into a comprehensive gesture force parameter C3. For example, a weighted summation method can be used, and the weight can be set to wi, C3 = sum(wi*pi), where i ranges from 1 to k. These gesture type codes C1, gesture motion trajectories C2, and gesture force parameters C3 together constitute the gesture behavior sub-feature D1 in the interactive behavior feature.

[0129] Step S143: extracting the gaze point coordinate sequence, gaze duration and pupil diameter change curve from the eye tracking data to generate the visual attention sub-feature in the interactive behavior feature.

[0130] Extract the gaze point coordinate sequence E1 from the eye tracking data B2. This coordinate sequence, recorded when the eye tracking data B2 was captured, consists of the two-dimensional or three-dimensional coordinates (xt, yt) or (xt, yt, zt) of the user's gaze point at each moment t. For ease of subsequent analysis, the coordinates may need to be normalized. Assuming the virtual scene space extends from x_min to x_max in the x direction, from y_min to y_max in the y direction, and from z_min to z_max in the z direction (if two-dimensional, only the x and y directions are considered), for the coordinates (xt, yt, zt), the normalized coordinates (xn_t, yn_t, zn_t) are calculated as follows: xn_t = (xt-x_min) / (x_max-x_min), yn_t = (yt-y_min) / (y_max-y_min), and zn_t = (zt-z_min) / (z_max-z_min) (for two-dimensional, only xn_t and yn_t are calculated). The normalized coordinate sequence is the extracted gaze point coordinate sequence E1, which more intuitively reflects the relative position of the gaze point in the scene.

[0131] Extract the gaze duration E2. The gaze duration E2 is recorded in the eye tracking data B2, that is, the length of time dt that the user stays at each gaze point. In order to better analyze the user's visual focus, it may be necessary to perform some statistical analysis on the gaze duration. For example, the entire interaction process is divided into several time periods, and the time periods are set as T1, T2, ..., Tn. The sum of the gaze durations of different gaze points in each time period is counted. Let the sum of the gaze durations of the gaze points (xj, yj, zj) in the time period Ti be Dij. By analyzing the values of these Dij, the degree of attention of the user to different areas in different time periods can be understood. These statistical gaze duration information constitute the gaze duration E2.

[0132] Extract the pupil diameter change curve E3. The pupil diameter change curve E3 also comes from the eye tracking data B2, that is, the curve formed by the change of the pupil diameter value pt at time t over time. In order to analyze the curve characteristics more clearly, it can be smoothed. Assuming that the moving average method is used, the window size is set to w. For the pupil diameter value pt at time t, the smoothed pupil diameter value p′t is calculated as: p′t = sum(pt-j) / w, where j ranges from 0 to w-1. The smoothed pupil diameter change curve E3 can more accurately reflect the changes in the user's visual attention intensity, because the pupil diameter is usually related to the user's attention and emotional state. The gaze point coordinate sequence E1, the gaze duration E2 and the pupil diameter change curve E3 jointly generate the visual attention sub-feature D2 in the interactive behavior feature.

[0133] Step S144: extracting speech text content, speech emotion polarity and speech instruction type from the speech input data to generate semantic parsing sub-features in the interactive semantic features.

[0134] Extract the speech text content F1 from the speech input data B3. This content is obtained during the initial processing of the speech signal, namely, the string S converted by speech recognition. For further semantic analysis, the speech text content may need to be segmented. Assume that a certain word segmentation algorithm is used to divide the string S into a series of words wi, where i ranges from 1 to 1. These words can then be tagged with parts of speech, such as nouns, verbs, and adjectives. Let the part-of-speech tagging result be ti, where i ranges from 1 to 1. This segmented and tagged speech text content constitutes the extracted speech text content F1, providing more detailed information for subsequent semantic understanding.

[0135] Extract the speech emotion polarity F2. The speech emotion polarity F2 has been obtained through sentiment analysis technology when processing the speech input data B3, that is, the numerical value e representing the positive, negative or neutral emotion in the speech. In order to make this value more in line with the needs of subsequent analysis, it may be necessary to normalize it. Assuming that the value range of the emotion polarity is from -1 to 1, map it to a new range, for example, from 0 to 1. Let the normalized speech emotion polarity be e', and the calculation method is: e'=(e+1) / 2. The normalized speech emotion polarity F2 can be more conveniently analyzed in combination with other features.

[0136] Extract the voice instruction type F3. The voice instruction type F3 has been determined when analyzing the voice text content and matching it with the preset instruction rules, that is, the identifier c representing different instruction types. In order to identify instructions more clearly, the instruction rules may need to be further refined. For example, for the "query material attributes" instruction, it can be further divided into querying different attributes of different types of materials. Suppose there are m types of materials and n types of attributes. For each instruction type c, it can be expanded to a combined identifier containing the material type and the attribute type. For example, c′=(mi, nj), where mi represents the i-th material type and nj represents the j-th attribute type. The refined voice instruction type F3 can more accurately reflect the user's operation intention. The voice text content F1, the voice emotional polarity F2 and the voice instruction type F3 jointly generate the semantic parsing sub-feature D3 in the interactive semantic feature.

[0137] Step S145: performing feature fusion on the gesture behavior sub-feature, the visual attention sub-feature, and the semantic analysis sub-feature to generate a complete interactive behavior feature and interactive semantic feature of the interactive operation instruction.

[0138] Step S1451: normalizing the spatial coordinates of the gesture motion trajectory in the gesture behavior sub-feature to obtain a normalized gesture trajectory.

[0139] For the gesture trajectory C2 in the gesture behavior sub-feature D1, spatial coordinate normalization is performed. Assume that the coordinates in the gesture trajectory C2 are the coordinates (x_t, y_t, z_t) in a three-dimensional space with the virtual reality device as the reference frame. The virtual scene space ranges from x_min to x_max in the x direction, from y_min to y_max in the y direction, and from z_min to z_max in the z direction. For each coordinate point (x_t, y_t, z_t), the normalized coordinates (xn_t, yn_t, zn_t) are calculated as follows: xn_t = (x_t - x_min) / (x_max - x_min), yn_t = (y_t - y_min) / (y_max - y_min), and zn_t = (z_t - z_min) / (z_max - z_min). This calculation yields the normalized gesture trajectory G1, which unifies the gesture trajectory into a relative spatial range, facilitating comparison and fusion with other features.

[0140] Step S1452: performing time window segmentation on the gaze point coordinate sequence in the visual attention sub-feature to obtain gaze area heat maps corresponding to multiple time segments.

[0141] The gaze point coordinate sequence E1 in the visual attention sub-feature D2 is segmented into time windows. Assume that the time range of the entire interaction process is divided into several time windows, each with a length of Δt. For each time window Ti, the gaze point coordinates (xt, yt, zt) within that time period are collected. Based on the virtual scene space, it is divided into several small regions, each with a range of Δx × Δy × Δz. The number of fixations within each small region is counted, with the number of fixations within a small region (xi, yi, zi) being ni. Based on this information, a gaze area heat map corresponding to each time window is generated. For example, the number of fixations ni can be converted into color values using a color mapping relationship. Let the color mapping function be map(ni). For each small region (xi, yi, zi), the corresponding color value is obtained through the map function based on its ni value, thereby plotting the gaze area heat map H1. In this way, the gaze area heat map H1 corresponding to multiple time segments obtained through time window segmentation can more intuitively reflect the distribution of the user's attention to different areas of the virtual scene during different time periods.

[0142] Step S1453: extract keywords from the speech text content in the semantic analysis sub-feature to obtain a semantic keyword set and keyword weight distribution.

[0143] Keyword extraction is performed on the spoken text content F1 in the semantic analysis sub-feature D3. First, the segmented and part-of-speech tagged spoken text content (words wi and parts of speech ti, where i ranges from 1 to 1) is analyzed. Keyword extraction algorithms, such as those based on term frequency-inverse document frequency (TF-IDF), can be used to determine the importance of each word by analyzing its frequency of occurrence in the spoken text and its inverse document frequency in a preset text collection (e.g., a collection of various textile culture-related texts). The importance score calculation function for word wi is set as score(wi), and the score of each word is calculated using this function. A threshold is then set, and words with scores above this threshold are considered semantic keywords. Let the semantic keyword set be K1, which contains keywords ki, where i ranges from 1 to s. Furthermore, a weight is assigned to each keyword based on score(wi), with the weight of keyword ki being wi', forming a keyword weight distribution K2. The semantic keyword set K1 and the keyword weight distribution K2 can highlight key information in the spoken text content and help more accurately understand the user's semantic intent.

[0144] Step S1454: Determine the association weight between gesture and visual attention based on the spatial overlap between the normalized gesture trajectory and the gaze area heat map.

[0145] Calculate the spatial overlap between the normalized gesture trajectory G1 and the gaze area heat map H1. For the normalized gesture trajectory G1, compare its position in the virtual scene space with the virtual scene area corresponding to the gaze area heat map H1. Assume that the virtual scene space is divided into smaller grid cells, and set the size of each grid cell to δx×δy×δz. Count the number of grid cells that the normalized gesture trajectory G1 passes through and the grid cells covered by the high-attention area in the gaze area heat map H1 (for example, the area where the number of gaze points exceeds a set threshold). Let the set of grid cells passed by the normalized gesture trajectory G1 be M1, the set of grid cells covered by the high-attention area in the gaze area heat map H1 be M2, and the number of overlapping grid cells be o. The spatial overlap calculation function is overlap(Ml, M2), and the spatial overlap value o′ is obtained through this function. Determine the association weight L1 between gesture and visual attention based on the spatial overlap value o′. For example, a function f(o′) can be set so that the association weight L1 = f(o′). The function can be adjusted according to actual conditions. For example, when the spatial overlap is higher, the association weight is larger to reflect the closeness between the gesture action and visual attention.

[0146] Step S1455: Determine the matching degree between the semantic instruction and the gesture and visual attention content according to the semantic keyword set and the keyword weight distribution.

[0147] The semantic keyword set K1 and keyword weight distribution K2 are used to determine the degree of content matching between semantic instructions and gestures and visual attention. For each keyword ki in the semantic keyword set K1, its correlation with the content related to the gesture behavior sub-feature D1 and the visual attention sub-feature D2 is analyzed. For example, if the keyword ki is "spinning wheel," the gesture motion trajectory C2 in the gesture behavior sub-feature D1 is related to the action of operating a spinning wheel, and the gaze area heat map H1 in the visual attention sub-feature D2 has a high degree of attention in the spinning wheel model area, then the keyword has a high correlation with the gesture and visual attention. A correlation analysis function relate(ki, D1, D2) is used to calculate the correlation score ri of each keyword with gesture and visual attention. Then, based on the weights wi' in the keyword weight distribution K2, the content matching degree M1 between the semantic instruction and the gesture and visual attention is calculated. For example, a weighted summation method can be used: M1 = sum(wi'*ri), where i ranges from 1 to s. This content matching degree M1 reflects the degree of consistency between the semantic instruction and the content of attention expressed by the user through gestures and visuals.

[0148] Step S1456: Based on the association weight and the content matching degree, the gesture behavior sub-feature, the visual attention sub-feature and the semantic analysis sub-feature are weightedly fused to generate the interaction behavior feature and the interaction semantic feature.

[0149] According to the association weight L1 and the content matching degree M1, the gesture behavior sub-feature D1, the visual attention sub-feature D2 and the semantic analysis sub-feature D3 are weightedly fused. Assume that the weighting coefficient of the gesture behavior sub-feature D1 is α, the weighting coefficient of the visual attention sub-feature D2 is β, and the weighting coefficient of the semantic analysis sub-feature D3 is γ. The values of α, β, and γ can be determined by the set functional relationship based on the association weight L1 and the content matching degree M1. For example, α = L1*M1, β = (1-L1)*M1, γ = 1-α-β (this is only an example functional relationship, which can be adjusted according to the actual situation). Then, the gesture behavior sub-feature D1, the visual attention sub-feature D2 and the semantic analysis sub-feature D3 are fused according to the weighting coefficients. Assume that the gesture behavior sub-feature D1 can be represented as a feature vector V1, the visual attention sub-feature D2 as a feature vector V2, and the semantic analysis sub-feature D3 as a feature vector V3. The fused feature vector V = α*V1+β*V2+γ*V3 (the "+" here represents vector concatenation, not numerical addition). This fused feature vector V constitutes the complete interactive behavior feature and interactive semantic feature N1 of the interactive operation instruction. It integrates the information expressed by the user through gestures, vision, and voice, and more comprehensively reflects the user's interactive intention.

[0150] Step S150: adjusting the three-dimensional model display parameters of the virtual reality textile culture scene according to the interactive behavior characteristics and the interactive semantic characteristics, generating an updated virtual reality textile culture scene and performing real-time rendering output.

[0151] Based on the complete interactive behavior features and interactive semantic features N1 of the obtained interactive operation instructions, the three-dimensional model display parameters of the virtual reality textile culture scene Z are adjusted to generate an updated virtual reality textile culture scene Z′ and render it in real time.

[0152] Step S151: Identify the model operation request type in the interactive behavior feature and the semantic adjustment instruction in the interactive semantic feature.

[0153] The interaction behavior features and interaction semantic features N1 are analyzed to identify the model operation request type O1 and semantic adjustment instruction O2.

[0154] For model operation request type O1, start from the interactive behavior feature part. For example, observe the gesture type code in the gesture behavior sub-feature. If the gesture type code matches the preset gesture code for model rotation, then the model operation request type is determined to be a model rotation request. Assuming that the preset model rotation gesture code set is Set1, when the gesture type code belongs to Set1, it is determined to be a model rotation request. Similarly, if the gesture type code matches the preset gesture code set Set2 for model scaling, it is determined to be a model scaling request.

[0155] The semantic adjustment instruction O2 is identified based on the voice instruction type in the semantic parsing sub-feature. For example, if the identifier corresponding to the voice instruction type is the same as the preset instruction identifier for querying material properties, assuming that identifier is ID1, when the voice instruction type is ID1, the semantic adjustment instruction is determined to be a material property query request. If the voice instruction type is the same as the preset instruction identifier ID2 for playing historical background explanation, it is determined to be a historical background playback request.

[0156] Step S152: When the model operation request type is a model rotation request, a rotation angle increment is calculated according to the gesture motion trajectory in the interactive behavior feature, and the Euler angle parameters of the corresponding three-dimensional model are adjusted.

[0157] When the model operation request type O1 is determined to be a model rotation request, the rotation angle increment P1 is calculated based on the gesture motion trajectory C2 (normalized gesture trajectory G1 after normalization) in the interactive behavior characteristics. First, analyze the changes in the gesture motion trajectory G1 in three-dimensional space. The calculation is based on the position changes of the starting point and end point of the gesture in space. Assume that the coordinates of the gesture starting point in three-dimensional space are (x_start, y_start, z_start), and the coordinates of the end point are (x_end, y_end, z_end). Through the calculation method of the spatial vector, the vector V1 from the starting point to the end point is obtained as (x_end-x_start, y_end-y_start, z_end-z_start).

[0158] At the same time, a reference vector V_ref is determined based on the initial orientation of the model in the virtual scene (for example, if the model is initially along the positive x-axis, V_ref = (1, 0, 0)). By calculating the angle between vector V1 and the reference vector V_ref (using the vector angle calculation method), an initial angle value θ1 is obtained. However, since gesture motion may contain some noise or irregularities, this angle value needs to be corrected. A correction function correct(θ1) can be set to correct θ1 and obtain the final rotation angle increment P1.

[0159] After obtaining the rotation angle increment P1, adjust the Euler angle parameters of the corresponding three-dimensional model. Assuming that the current Euler angle parameters of the three-dimensional model are (Euler_x, Euler_y, Euler_z), according to the rotation angle increment P1 and the setting of the rotation axis (for example, if it is rotating around the y-axis), the new Euler angle parameters are calculated as follows: If rotating around the y-axis, the new Euler_y = Euler_y + P1, while Euler_x and Euler_z remain unchanged (here only the rotation around the y-axis is taken as an example. In practice, it is possible to rotate around different axes according to different needs, and the calculation method is adjusted accordingly). In this way, the adjustment of the Euler angle parameters of the corresponding three-dimensional model is completed to realize the rotation operation of the model.

[0160] Step S153: When the model operation request type is a model scaling request, a scaling coefficient is calculated according to the gesture force parameter in the interaction behavior feature, and vertex coordinate parameters of the corresponding three-dimensional model are adjusted.

[0161] When the model operation request type O1 is a model scaling request, the scaling coefficient Q1 is calculated based on the gesture force parameter C3 in the interactive behavior characteristics. The gesture force parameter C3 is obtained by comprehensively calculating the pressure values of various parts of the hand. Assume that the value range of the gesture force parameter C3 is [C3_min, C3_max], and the value range of the scaling coefficient Q1 is set to [Q1_min, Q1_max]. The gesture force parameter C3 is mapped to the value range of the scaling coefficient Q1 through a mapping function map(C3). For example, the mapping function can be a linear function, which determines the corresponding value of C3 in [Q1_min, Q1_max] according to its position in [C3_min, C3_max], thereby obtaining the scaling coefficient Q1.

[0162] After obtaining the scaling factor Q1, adjust the vertex coordinate parameters of the corresponding 3D model. Assume that the vertex coordinate set of the 3D model is Set_V, where each vertex coordinate is (xi, yi, zi), and i ranges from 1 to n (n is the number of vertices). For each vertex coordinate (xi, yi, zi), the new vertex coordinates (xi_new, yi_new, zi_new) are calculated as follows: xi_new = xi*Q1, yi_new = yi*Q1, zi_new = zi*Q1. By performing this scaling calculation on each vertex coordinate, the corresponding 3D model vertex coordinate parameters are adjusted, achieving the model scaling operation.

[0163] Step S154: When the semantic adjustment instruction includes a material attribute query request, the corresponding textile material attribute information is extracted from the textile culture data unit, and a material attribute annotation layer is superimposed on the surface of the three-dimensional model.

[0164] When semantic adjustment instruction O2 includes a material attribute query request, the corresponding textile material attribute information V is extracted using textile culture data unit Y as the data source. Assume that the textile material attribute information V in textile culture data unit Y is stored in a data structure, such as a dictionary containing various attributes and their corresponding values. Let the attribute name be attr_name and the corresponding value be attr_value. The dictionary can be represented as Dict = {attr_name1:attr_value1, attr_name2:attr_value2, ...}.

[0165] Based on the attribute name specified in the query request (e.g., the attribute name to be queried is determined from the semantic keyword set K1), the corresponding attribute value is extracted from the dictionary Dict. For example, if the query request is for the attribute "fiber length", the attr_value corresponding to attr_name "fiber length" is found in the dictionary.

[0166] After extracting the corresponding textile material attribute information, a material attribute annotation layer is overlaid on the 3D model surface. Assume the 3D model is textile material 3D model Y. First, determine the appropriate location on the 3D model surface to overlay the annotation layer. This can be determined based on the 3D model's structure and the user's current viewing angle. For example, if the user's current viewing angle is facing a specific surface of the model, select a location relative to the center of that surface as the annotation starting point.

[0167] Assume that the content of the annotation layer is text information containing attribute names and values, such as "Fiber length: at tr_value." This text information is displayed at a specific location on the 3D model surface using a specified font, color, and size. The font is determined by a font parameter (Font_param), the color by a color parameter (Color_param), and the size by a size parameter (Size_param). In this way, a material attribute annotation layer is superimposed on the 3D model surface, allowing users to intuitively obtain the required material attribute information.

[0168] Step S155: When the semantic adjustment instruction includes a historical background playback request, the corresponding textile history information is extracted from the textile culture data unit, and a historical explanation audio stream is inserted into the virtual reality textile culture scene.

[0169] When the semantic adjustment instruction O2 includes a request to play historical background information, the corresponding textile historical information W is extracted from the textile cultural data unit Y. Textile historical information W may include multiple aspects, such as origin information and development stage information. Assume that this information is stored in a data structure, such as a list List, where each element represents a piece of historical information, such as List = [info1, info2, ...]. Info1 may describe the origin time and background, while info2 may describe the technological changes at a certain development stage.

[0170] According to the specific requirements of the historical background playback request (for example, determining the historical information theme to be played from the semantic keyword set K1), the corresponding historical information segment is extracted from the list List. For example, if the request is to play "origin history", the corresponding origin information segment infol is found in the list.

[0171] After extracting the corresponding textile history information, we insert the historical explanation audio stream into the virtual reality textile culture scene Z. First, we prepare the historical explanation audio file. Assume that the audio file is stored in a designated directory and the file name corresponds to the historical information segment. For example, the audio file corresponding to info1 is named audio1.mp3.

[0172] In VR Textile Culture Scene Z, determine the playback location and mode for the audio stream. The playback location can be determined based on the scene's layout and the user's current position. For example, placing the audio source close to the user's viewpoint can enhance the user's auditory experience. The playback mode can be set to automatic or triggered based on further user actions.

[0173] Set the audio stream's volume parameter to Volume_param. Adjusting this parameter controls the audio playback volume. You can also set the audio playback mode, such as loop or single play, using the Play_mode_param parameter. With this setup, you can insert a historical explanatory audio stream into the VR textile culture scene, providing users with historical background information.

[0174] Step S156: performing scene synchronization update on the adjusted three-dimensional model display parameters and the unadjusted three-dimensional model to generate the updated virtual reality textile culture scene.

[0175] Step S1561: Detecting display status change marks of each three-dimensional model in the virtual reality textile culture scene.

[0176] In the virtual reality textile culture scene Z, a display status change flag Flag is set for each three-dimensional model. When the display parameters of the three-dimensional model (such as Euler angle parameters, vertex coordinate parameters, etc.) are adjusted, the corresponding display status change flag Flag is set to "adjusted"; if no adjustment occurs, it is set to "unadjusted". For example, for the textile process three-dimensional model J, if its Euler angle parameters change due to a model rotation request, J's display status change flag Flag_J is set to "adjusted"; if the parameters of the historical textile scene three-dimensional model S do not change, Flag_S is set to "unadjusted". By detecting the display status change flag of each three-dimensional model, it is clear which models' parameters have changed and which models remain unchanged.

[0177] Step S1562: For the three-dimensional model whose display state change flag is adjusted, the position matrix of the model in the virtual space is recalculated using a quaternion interpolation algorithm according to the updated Euler angle parameters or vertex coordinate parameters.

[0178] For a 3D model whose display status change flag is "Adjusted," take the updated Euler angle parameters after the model is rotated as an example (the same applies to updating vertex coordinate parameters when the model is scaled). Assume that the adjusted 3D model is a textile process 3D model J, and its updated Euler angle parameters are (Euler_x_new, Euler_y_new, Euler_z_new).

[0179] A quaternion interpolation algorithm is used to recalculate its position matrix M in virtual space. First, the Euler angle parameters are converted to quaternion representation. Let the conversion function be euler_to_quaternion(Euler_x_new, Euler_y_new, Euler_z_new), and the corresponding quaternion q = (q0, q1, q2, q3) is obtained through this function.

[0180] The quaternion interpolation algorithm requires determining the starting and ending states of the interpolation. Assume that the quaternion of the model before adjustment is represented by q_start = (q0_start, q1_start, q2_start, q3_start), and the updated quaternion is q_end = (q0, q1, q2, q3). Let the interpolation time parameter be t, with a value range of [0, 1], representing the degree of transition from the starting state to the ending state.

[0181] The interpolation quaternion q_interp=interpolate(q_start, q_end, t) under the time parameter t is calculated using the quaternion interpolation formula.

[0182] Finally, the interpolated quaternion q_interp is converted back to the position matrix M. Let the conversion function be quate mion_to_matrix(q_interp), through which the position matrix M in the virtual space is obtained. The matrix determines the new position and posture of the 3D model in the virtual space to achieve smooth transition and accurate spatial positioning.

[0183] Step S1563: For the three-dimensional model whose display status change flag is not adjusted, keep its original position matrix unchanged.

[0184] For 3D models whose display status change is marked as "unadjusted," such as the 3D model S in the historical textile scene, their original position matrix M_original remains unchanged, as their display parameters remain unchanged. This is because during the scene synchronization update process, the position matrix of unadjusted models does not need to be recalculated, reducing the computational effort and maintaining a stable relative position between models within the scene. In this way, within the entire VR textile culture scene, 3D models in different states can collaboratively form a unified scene structure that meets the user's interaction intent.

[0185] Step S1564: Reconstruct the spatial topology structure of the virtual reality scene based on the position matrices of all three-dimensional models.

[0186] Based on the position matrices of all three-dimensional models, including the recalculated position matrices of the adjusted models and the original position matrices of the unadjusted models, the spatial topology of the virtual reality scene is reconstructed.

[0187] Let Set_models be the set of all 3D models, where each 3D model has its corresponding position matrix. For example, the position matrix of textile process 3D model J is M_J, the position matrix of historical textile scene 3D model S is M_S, the position matrix of textile material 3D model Y is M_Y, and so on.

[0188] The spatial topology of a virtual reality scene describes the relative positions and connections of the various 3D models in the virtual space. The spatial relationships between the models are determined by analyzing the information in each 3D model's position matrix, such as its translation, rotation, and scaling parameters. For example, if the translation parameter in one model's position matrix indicates a greater x-axis offset than another model, then in the spatial topology, that model is relatively further away from the reference point on the x-axis.

[0189] At the same time, possible hierarchical or associative relationships between models should be considered. For example, a 3D textile material model Y may be part of a component within a 3D textile process model J. This relationship also needs to be reflected in the spatial topology. By integrating the position matrix information of all 3D models, a spatial topology that accurately reflects the spatial layout and relationships of the models in the current scene is reconstructed.

[0190] Step S1565: re-rendering all three-dimensional models in the virtual reality textile culture scene based on the updated spatial topological structure, so that the relative positional relationship between the three-dimensional models is consistent with the expected effect after the user interaction operation.

[0191] Based on the updated spatial topology, all 3D models in the virtual reality textile culture scene are re-rendered. The rendering process involves multiple aspects, including the model's geometry, texture mapping, lighting effects, etc.

[0192] For each 3D model, its display position in the virtual scene is determined based on its position and posture in the spatial topology. For example, the model is placed in the correct position in the virtual space and its orientation is adjusted using the translation and rotation parameters in the position matrix.

[0193] In terms of texture mapping, ensure that the textures previously applied to the 3D model (such as the texture mapping relationship of the textile material 3D model, the operation trajectory texture of the textile process 3D model, etc.) can be accurately displayed on the updated model surface. This may require fine-tuning the texture UV coordinates based on the new pose and position of the model to ensure texture continuity and accuracy.

[0194] Lighting also needs to be adjusted based on the spatial topology and the model's new position. For example, if a model's position changes, the angle and intensity of the ambient and direct light it receives will also change. By recalculating lighting parameters such as intensity, color, and direction, the model's surface will appear lit in a manner consistent with the actual scene. For example, if a model is rotated, the distribution of highlights and shadows on its surface needs to be adjusted accordingly to simulate realistic lighting conditions.

[0195] By re-rendering all 3D models in terms of geometry, texture mapping, and lighting effects, the relative positions of the 3D models are ensured to align with the intended effect after user interaction. This allows users to see the virtual reality textile culture scene in a manner consistent with their intended operation, completing the entire process from user interaction to scene update and rendering. Finally, the updated virtual reality textile culture scene is rendered and output in real time, presenting users with a dynamic and interactive textile culture display.

[0196] Figure 2 A schematic diagram illustrates exemplary hardware and software components of a virtual reality-based textile cultural knowledge visualization system 100, provided in some embodiments of the present application, that can implement the concepts of the present application. For example, a processor 120 can be used in the virtual reality-based textile cultural knowledge visualization system 100 to perform the functions described in the present application.

[0197] The textile cultural knowledge visualization presentation system 100 combined with virtual reality can be a general-purpose server or a special-purpose server, both of which can be used to implement the textile cultural knowledge visualization presentation method combined with virtual reality of this application. Although this application only shows a single server, for convenience, the functions described in this application can be implemented in a distributed manner on multiple similar platforms to balance the processing load.

[0198] For example, the textile cultural knowledge visualization presentation system 100 combined with virtual reality may include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and storage media 140 in different forms, such as a disk, ROM, or RAM, or any combination thereof. Exemplarily, the textile cultural knowledge visualization presentation system 100 combined with virtual reality may also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The method of the present application can be implemented according to these program instructions. The textile cultural knowledge visualization presentation system 100 combined with virtual reality also includes an I / O interface 150 between the computer and other input and output devices.

[0199] For ease of explanation, only one processor is described in the textile culture knowledge visualization presentation system 100 combined with virtual reality. However, it should be noted that the textile culture knowledge visualization presentation system 100 combined with virtual reality in the present application may also include multiple processors, so the steps performed by one processor described in the present application may also be performed jointly or individually by multiple processors. For example, if the processor of the textile culture knowledge visualization presentation system 100 combined with virtual reality executes step A and step B, it should be understood that step A and step B may also be performed jointly by two different processors or individually in one processor. For example, the first processor executes step A, the second processor executes step B, or the first processor and the second processor execute steps A and B together.

[0200] In addition, an embodiment of the present invention further provides a readable storage medium, in which computer-executable instructions are preset. When a processor executes the computer-executable instructions, the above-mentioned method for visualizing textile cultural knowledge combined with virtual reality is implemented.

[0201] It should be noted that in order to simplify the description of the present invention and thus help understand one or more embodiments of the invention, in the foregoing description of the embodiments of the present invention, multiple features are sometimes combined into one embodiment, figure or description thereof.

Claims

1. A method for visualizing textile cultural knowledge combined with virtual reality, characterized in that: The method comprises: Acquire a target textile culture data set, wherein the target textile culture data set includes a plurality of textile culture data units, each of which includes textile process information, textile history information, and textile material attribute information; constructing a three-dimensional textile model set based on the target textile culture data set, wherein the three-dimensional textile model set includes three-dimensional models of multiple textile cultural elements; According to preset virtual reality display rules, each three-dimensional model in the three-dimensional textile model set is dynamically matched with the corresponding textile culture data unit to generate a virtual reality textile culture scene; Responding to the user's interactive operation instructions in the virtual reality textile culture scene, extracting interactive behavior features and interactive semantic features in the interactive operation instructions; The three-dimensional model display parameters of the virtual reality textile culture scene are adjusted according to the interactive behavior characteristics and the interactive semantic characteristics, and an updated virtual reality textile culture scene is generated and rendered and output in real time.

2. The method for visually presenting textile cultural knowledge combined with virtual reality according to claim 1 is characterized in that: The constructing of a three-dimensional textile model set based on the target textile culture data set includes: Parsing each textile culture data unit in the textile culture data set to obtain parsing results; the parsing results include a textile process step sequence, a textile history timeline, and physical property parameters of textile materials; extracting textile process characteristics, textile history association characteristics and textile material structure characteristics from the analysis results; Input the textile process characteristics into a preset three-dimensional process model generation module to generate a corresponding textile process three-dimensional model; input the textile history association characteristics into a preset historical scene generation module to generate a corresponding historical textile scene three-dimensional model; input the textile material structure characteristics into a preset material structure analysis module to generate a textile material three-dimensional model; According to preset textile cultural element matching rules, the textile process three-dimensional model, the historical textile scene three-dimensional model and the textile material three-dimensional model are combined to obtain each three-dimensional model in the three-dimensional textile model set.

3. The method for visualizing textile cultural knowledge combined with virtual reality according to claim 2 is characterized in that: The combining of the textile process three-dimensional model, the historical textile scene three-dimensional model, and the textile material three-dimensional model according to a preset textile cultural element matching rule includes: Obtaining a process time label corresponding to the textile process three-dimensional model, a historical period label corresponding to the historical textile scene three-dimensional model, and a material application period label corresponding to the textile material three-dimensional model; Determining the spatiotemporal matching degree between the three-dimensional model of the textile process, the three-dimensional model of the historical textile scene, and the three-dimensional model of the textile material according to the time alignment relationship between the process time label, the historical period label, and the material application period label; When the spatiotemporal matching degree exceeds a preset matching threshold, the corresponding textile process three-dimensional model, the historical textile scene three-dimensional model, and the textile material three-dimensional model are hierarchically combined to form a composite three-dimensional model including a process operation layer, a historical scene layer, and a material structure layer; When the spatiotemporal matching degree does not exceed the matching threshold, the textile process three-dimensional model, the historical textile scene three-dimensional model and the textile material three-dimensional model are respectively added to the three-dimensional textile model set as independent three-dimensional models.

4. The method for visualizing textile cultural knowledge combined with virtual reality according to claim 1 is characterized in that: The method of dynamically matching each three-dimensional model in the three-dimensional textile model set with a corresponding textile culture data unit according to a preset virtual reality display rule to generate a virtual reality textile culture scene includes: Extracting cultural display priority parameters, user preference matching parameters, and device rendering performance parameters from the textile culture data unit; determining an initial display position of each three-dimensional model in the three-dimensional textile model set according to the cultural display priority parameter; Adjusting the display size, display angle, and display dynamic effect of the three-dimensional model according to the user preference matching parameters; Optimizing the texture resolution, number of model faces, and lighting calculation complexity of the three-dimensional model according to the device rendering performance parameters; The adjusted three-dimensional models are arranged according to preset spatial layout rules, and virtual reality environment lighting effects and background sound effects are added to generate the virtual reality textile culture scene.

5. The method for visually presenting textile cultural knowledge combined with virtual reality according to claim 4 is characterized in that: The adjusting the display size, display angle, and display dynamic effect of the three-dimensional model according to the user preference matching parameters includes: Identifying historical interaction record features, user attention distribution features, and user device operation habit features in the user preference matching parameters; Adjusting the display size of the three-dimensional model according to the historical interaction record characteristics, so that the display size of the three-dimensional model with a higher historical interaction frequency of the user is larger; Adjusting the display angle of the three-dimensional model according to the user attention distribution characteristics so that the viewing angle of the three-dimensional model with higher user attention is closer to the user's viewpoint center area; The dynamic display effect of the three-dimensional model is adjusted according to the user device operation habit characteristics, so that the rotation speed and zoom sensitivity of the three-dimensional model match the operation delay parameters in the user device operation habit characteristics.

6. The method for visualizing textile cultural knowledge combined with virtual reality according to claim 1 is characterized in that: The step of responding to the user's interactive operation instructions in the virtual reality textile culture scene and extracting interactive behavior features and interactive semantic features in the interactive operation instructions includes: Capture user gesture data, eye tracking data, and voice input data in virtual reality devices: Extracting gesture type codes, gesture motion trajectories, and gesture force parameters from the gesture action data to generate gesture behavior sub-features in the interaction behavior features; extracting a gaze point coordinate sequence, a gaze duration, and a pupil diameter change curve from the eye tracking data to generate a visual attention sub-feature in the interactive behavior feature; Extracting speech text content, speech emotion polarity, and speech command type from the speech input data to generate semantic parsing sub-features in the interactive semantic features; The gesture behavior sub-feature, the visual attention sub-feature, and the semantic analysis sub-feature are fused to generate a complete interaction behavior feature and interaction semantic feature of the interaction operation instruction.

7. The method for visually presenting textile cultural knowledge combined with virtual reality according to claim 6 is characterized in that: The feature fusion of the gesture behavior sub-feature, the visual attention sub-feature, and the semantic analysis sub-feature includes: Normalizing the spatial coordinates of the gesture motion trajectory in the gesture behavior sub-feature to obtain a normalized gesture trajectory; Performing time window segmentation on the gaze point coordinate sequence in the visual attention sub-feature to obtain gaze area heat maps corresponding to multiple time segments; Extract keywords from the speech text content in the semantic analysis sub-feature to obtain a semantic keyword set and keyword weight distribution; determining an association weight between gesture and visual attention based on a spatial overlap between the normalized gesture trajectory and the gaze area heat map; Determining a degree of matching between the semantic instruction and the gesture and visual attention content based on the semantic keyword set and the keyword weight distribution; The gesture behavior sub-feature, the visual attention sub-feature, and the semantic analysis sub-feature are weightedly fused based on the association weight and the content matching degree to generate the interaction behavior feature and the interaction semantic feature.

8. The method for visually presenting textile cultural knowledge combined with virtual reality according to claim 1 is characterized in that: The adjusting the three-dimensional model display parameters of the virtual reality textile culture scene according to the interactive behavior characteristics and the interactive semantic characteristics to generate an updated virtual reality textile culture scene includes: Identifying a model operation request type in the interactive behavior feature and a semantic adjustment instruction in the interactive semantic feature; When the model operation request type is a model rotation request, calculating the rotation angle increment according to the gesture motion trajectory in the interactive behavior feature, and adjusting the Euler angle parameters of the corresponding three-dimensional model; When the model operation request type is a model scaling request, a scaling coefficient is calculated according to the gesture force parameter in the interaction behavior feature, and vertex coordinate parameters of the corresponding three-dimensional model are adjusted; When the semantic adjustment instruction includes a material attribute query request, extracting corresponding textile material attribute information from the textile culture data unit, and superimposing a material attribute annotation layer on the surface of the three-dimensional model; When the semantic adjustment instruction includes a historical background playback request, extracting corresponding textile history information from the textile culture data unit and inserting a historical explanation audio stream into the virtual reality textile culture scene; The adjusted three-dimensional model display parameters and the unadjusted three-dimensional model are synchronously updated to generate the updated virtual reality textile culture scene.

9. The method for visually presenting textile cultural knowledge combined with virtual reality according to claim 8, characterized in that: The step of synchronously updating the adjusted three-dimensional model display parameters with the unadjusted three-dimensional model to generate the updated virtual reality textile culture scene includes: Detecting a display status change mark of each three-dimensional model in the virtual reality textile culture scene; For 3D models whose display status change flag is adjusted, the quaternion interpolation algorithm is used to recalculate their position matrix in the virtual space based on their updated Euler angle parameters or vertex coordinate parameters; For the 3D model whose display status change flag is not adjusted, its original position matrix is kept unchanged; Reconstruct the spatial topology of the virtual reality scene based on the position matrix of all three-dimensional models; All three-dimensional models in the virtual reality textile culture scene are re-rendered based on the updated spatial topological structure, so that the relative positional relationship between the three-dimensional models is consistent with the expected effect after the user interaction operation.

10. A textile cultural knowledge visualization presentation system combined with virtual reality, characterized in that: It includes a processor and a memory, the memory is connected to the processor, the memory is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the memory to implement the method for visual presentation of textile cultural knowledge combined with virtual reality as described in any one of claims 1 to 9.

Images